With the improvement of industrial automation, the demand for online real-time measurement of workpieces is growing. It is of great significance to detect the difference between the three-dimensional morphology of the measured workpiece and the standard workpiece, so as to realize the rapid screening of qualified products. The traditional 3D difference detection method usually adopts the reverse fringe technology. It is based on the coordinate correspondence between the camera and the projector, and the stripes projected on the standard workpiece are distributed in a straight line. When the workpiece to be measured is placed at the same position, the difference between the workpiece to be measured and the standard workpiece is directly reflected by the deformation of the straight stripe. In this paper, a real-time computer-generated moire profilometry(CGMP) for workpiece 3D difference detection is proposed. The single frame acquisition characteristics of CGMP are calculated. Only the wrapped phase is calculated and the difference evaluation function model is established. The size of the evaluation function is used to directly reflect the difference characteristics. The alignment position between the standard workpiece and the measured workpiece is tracked through the minimum point of the evaluation function, so as to achieve accurate detection of the difference between the two. The collected fringe data information has been saved in the process of difference detection. The real-time 3D reconstruction can be performed directly at the alignment position. The experimental results verify the feasibility of the proposed method in 3D difference detection.